This article is a radical re-writing of a paper given to a seminar at the World Bank in Washington in May 1982.

¹ Most of the south Indian fieldwork on which this article is based was conducted in 1977–78 in six villages in Anekal Taluk, Bangalore District, in the south Indian state of Karnataka. See Hill, Polly, Dry Grain Farming Families: Hausaland (Nigeria) and Karnataka (India) compared (Cambridge University Press, 1982).CrossRefGoogle Scholar

² ‘A thing given or granted; something known or assumed as fact, and made the basis of reasoning or calculation.’ Shorter Oxford Dictionary.

³ An idea derived from the following citation from C. A. Bayly: ‘What has happened is that in the absence of historical data a consensus has been built up on the basis of random insights …’. From Dewey, C. and Hopkins, A. G. (eds), The Imperial Impact (London, 1978), p. 172.Google Scholar

⁴ FAO, 1966.

⁵ Zarkovich is here citing Hendricks, W. A. in ‘Improving the Quality of Statistical Surveys’ (American Statistical Association, 1956), pp. 31–9.Google Scholar

⁶ Norman, D. W., ‘Crop Mixtures under Indigenous Conditions in the Northern part of Nigeria’, in Ofori, E. D. (ed.), Factors of Agricultural Growth in West Africa (University of Ghana, Legon, 1973), p. 130.Google Scholar

⁷ Of course these instructions must sometimes (or usually?) have been ignored. Thus, in other connections, different Indian states adopt different statistical methods when dealing with mixed cropping. See Bansil, P. S., Agricultural Statistics in India (Delhi, 1970).Google Scholar

⁸ Entitled ‘Note on rural surveys covering food consumption and household expenditure in tropical West Africa’, it is included in FAO Studies in Agricultural Economics and Statistics, 1952–1977, being selected articles from The Monthly Bulletin of Agricultural Economics and Statistics (FAO, Rome, 1978).Google Scholar

⁹ In the general field under discussion averages are almost always worthless–in the words of Lenin (see Hill, , Dry Grain Farming Families, p. 40Google Scholar) they are ‘fictitious’; but this particular average is the most breath-taking I have ever encountered.

¹⁰ But the concept is often meaningless, each event being an isolate.

¹¹ By Casely, D. J. and Lury, D. A. (Oxford, Clarendon Press, 1981).Google Scholar

¹² Delhi.

¹³ I write in the past tense since it seems that all the original initiative was taken by mathematicians and that the series has since become petrified.

¹⁴ In Dewey and Hopkins, The Imperial Impact.

¹⁵ for reference to this unpublished PhD thesis see Dewey, , in The Imperial Impact, p. 382, n. 6.Google Scholar

¹⁶ For reference see ibid., p. 383, n. 15.

¹⁷ Perhaps the most famous of all debates was that initiated in Current Anthropology (University of Chicago) by Harris, M., ‘The Cultural Ecology of India's Sacred Cattle’ (02 1966).Google Scholar

¹⁸ This extends from attempts to count cattle for myself to detailed analyses of published statistics. See Hill, Polly, Studies in Rural Capitalism in West Africa (Cambridge, 1970)Google Scholar, on Northern Ghanaian cattle censuses and other Ghanaian cattle statistics.

¹⁹ Stamp was incorrect in assuming that this lowly officiai, who was much inferior to the village accountant (patwari, shanbhog, etc., according to region), did this work in all areas. (See Dewey, , The Imperial Impact, p. 282).Google Scholar

²⁰ Some Economic Factors in Modem Life (London, 1929)Google Scholar, quoted in ibid., p. 281.

²¹ A Scheme for an Economic Census of India (Delhi, 1934).Google Scholar

²² The enumerator may well be unable to enter the house, as in Muslim Hausaland.

²³ Hill, Polly, Population, Prosperity and Poverty (Cambridge, 1977), Ch. XII.Google Scholar

²⁴ See Hill, Polly, ‘Joint Families in Rural Karnataka’, Modem Asian Studies, 14, 1 (1980).CrossRefGoogle Scholar

²⁵ Thus in Hausa villages male members of impoverished households may buy most of their cooked food from other households; and in many societies mothers provide much of their children's food.

²⁶ See Hill, , Dry Grain Farming Families.Google Scholar

²⁷ In Karnataka only the very richest farmers own separate cattle sheds, most animals living inside their owner's houses.

²⁸ The widespread use of indigenous units of ‘area’, based, for instance, on the quantity of grain required for sowing the plot, is, of course, a great complication.

²⁹ I make this dogmatic assertion on the basis of much detailed comparison of respondents' estimates of their holdings in the Karnataka villages and recorded land revenue statistics adjusted in accordance with respondents' own statements—relating, for instance, to unreported division of farmland between sons. It is because I know my statement to be correct, that I am so dismayed by the provocation it causes in certain academic circles.

³⁰ See Hill, , Dry Grain Farming Families, Appendix II (1). (Of course, I cannot say how representative Karnataka is of south India generally.)Google Scholar

³¹ Thus a father may deliberately make temporary allocations to his sons to prevent them selling the land.

³² A few villages were selected for resurvey following the census, each being the subject of a separate monograph.

³³ Thus in 1977 the ill-drawn, purchasable, mimeographed, official maps for the Anekal villages showing farm boundaries and survey numbers, were nearly half a century old, none of the subsequent (and numerous) farm-subdivisions being recorded. (Again, I am ignorant of the degree of representativeness of these maps, except that I doubt whether any more up-to-date maps were available for rural Bangalore District.)

³⁴ Hill, , Dry Grain Farming Families, p. 14.Google Scholar

³⁵ This may present grave difficulties, and the photographs may be out-of-date. (It is relevant to note that all printed Indian maps, other than those on a very small scale, were state secrets until recently—no post-Independence maps having been available, even in libraries; in 1981 it was announced that they would soon be issued except for areas within 50 miles of the sea.)

³⁶ The henna hedge boundaries of parts of Hausaland are ideal.

³⁷ Hill, Polly, Rural Hansa: A Village anda Setting (Cambridge, 1972)CrossRefGoogle Scholar and Population, Prosperity and Poverty.

³⁸ ‘Madras Government Statistics, 1880–1940’ in Baker, C. J. and Washbrook, D. A., Political Institutions and Political Change, 1880–1940 (Macmillan, Delhi, 1975).Google Scholar

³⁹ Report on the Timely Reporting Survey of Agricultural Statistics in Kerala 1977–78 (Bureau of Economics and Statistics, Government of Kerala, 1980).Google Scholar

⁴⁰ Not the least of the miserable investigator's difficulties is that in this area of dispersed settlement there are two types of ‘village’, which everyone, including the census authorities themselves, always confuses—neither of which has properly recognized boundaries.

⁴¹ Output is always measured by volume (or units, as with coconuts) not by weight. Paddy production may have to be measured if labourers are rewarded with a share of the crop harvested. As for share-cropping (which did not exist in Hausaland or Karnataka), this would usually apply to individual plots, aggregation being uninteresting.

⁴² The fact that each Karnataka grain measure, from the most diminutive to the largest, has a name, seems to authenticate it. despite the inefficiency of those who manufacture them.

⁴³ In passing, I insist that it is a general rule that official rural statistics, like prices in rural market-places, which are not collected for any evident practical purposes, are seldom checked by officials at any level—a fact which is only too obvious to the demoralized enumerator at the lowest level. Thus, village sheep and goat statistics, unlike some cattle statistics (cf. p. 497 above), are invariably useless, for veterinarians take little interest in small livestock: worst of all are fowls!

⁴⁴ Of course the concept of ‘average income’ is virtually meaningless in those tropical countries, such as Nigeria, with significant proportions of rich urban households.

⁴⁵ Fortes, Meyer, The Dynamics of Clanship among the Tallensi (London, Oxford University Press, 1945, p. x).Google Scholar

⁴⁶ But in certain circumstances the distinction may be very strict. Thus, the early migrant cocoa farmers of southern Ghana were not prepared to remunerate labourers from cash ear-marked for the purchase of land, but only from the proceeds of selling the cocoa; so they usually employed no labourers until the trees came into bearing. Hill, Polly, The Migrant Cocoa-Farmers of Southern Ghana (Cambridge, 1963).Google Scholar

⁴⁷ Especially in West Africa where very high proportions of wives engage in trade entirely independently of their husbands.

⁴⁸ It is because in many tropical countries (other than those which derive a large per capita income from exporting oil or manufactures) the bulk of the national income continues to be generated in the countryside, that this article on rural statistics properly includes a discussion of the ‘league table’ of per capita income by country.

⁴⁹ I use this vague word as a measure of ‘well-being’ rather than the misleading ‘income’. See Hill, , Dry Grain Farming Families.Google Scholar

⁵⁰ In practice creditors have to be identified by enquiry of debtors or other confidential informants: thus in the Anekal villages no one admitted to lending out a cow, though many admitted to borrowing one; and mortgagees and pledgees hardly ever declare themselves.

⁵¹ Thus, for example, many debts are only partially repaid—which effectively reduces the ‘debt’ if that had been the creditor's original expectation.

⁵² This interesting question (which, like the matter of aggregation, involves indigenous concepts of time) is seldom discussed—except very esoterically. There are many levels of difficulty, running from such humdrum aspects as the falsity of most written agreements and partial repayment, to the refusal of personal relationships to be regulated by uncomputable figures. (This is an unwritten book.)

⁵³ Their situation resembles that of Hausa slaves in early colonial times, who were omitted from the population censuses, which were based on tax registers, because they were not liable for tax. See Hill, , Population, Prosperity and Poverty, p. 201.Google Scholar

⁵⁴ Of course an acre of irrigated land may be well worth farming, but throughout this discussion I have had dry grain farmers in mind. (A very common defect of Indian land-holding statistics is the failure to distinguish (high-yielding) paddy land from dry land.) See Hill, , Dry Grain Farming Families.Google Scholar

⁵⁵ The 1981 census reports have not been published at the time of writing.

⁵⁶ It is entirely respectable to extrapolate from six villages only to Bangalore District villages in general, on the basis of the ratios of female labourers to the total female population in all the villages in the census report.

⁵⁷ Such grouping was quite unnecessary as each village had been enumerated separately.

⁵⁸ Whenever I have discussed this important subject publicly in India I have invariably failed to convince my listeners that the cause of the village-grouping was not dispersal of the original inhabitants of Hullahalli—in fact my evidence is that Hullahalli is not a peculiar case and that in south-eastern Karnataka the concept of breakaway villages does not apply. It is because the census-grouping makes no sociological or administrative sense, and is entirely unnecessary, that one's evidence is, unfortunately, disbelieved.

⁵⁹ This grouping has nothing to do with the recent creation of village panchayats, but arises simply from the rigid adherence to ancient village lists. (Hullahalli, including the two ‘stranger-villages’, is actually but a portion of Sakalavara Panchayat.)

⁶⁰ Statisticians are very much given to analysing village-size using the census figures; but doubts as to the reliability of the figures are never expressed.

⁶¹ Judging from my own humble enquiries in six Karnataka villages, and from the views I have heard generally expressed, I have no reason to doubt that the total figures (males and females) yielded by the census are other than reasonably reliable.

⁶² Ed. Farmer, B. H. and sub-titled Technology and Change in Rice-Crowing Areas of Tamil Nadu and Sri Lanka (Macmillan, London. 1977).Google Scholar

⁶³ When in the field with my assistants I have been constantly struck by their readiness to accept contradictory statements, which it was then my function to query; not to believe what is said, especially if the respondent is one's own countryman, is so embarrassingly rude that critical faculties are suspended.

⁶⁴ Whatever allowance is made for the extreme difficulty of defining ‘literacy’, it is certain that all censuses exaggerate its incidence; certainly, the proportion of rural householders capable of keeping records of (say) expenditure on every day for a month is very small indeed.